This letter is in response to the recent article in Journal of Clinical
Microbiology (38[2]:881-882, 2000) entitled "Failure to detect
Chlamydia pneumoniae in brain sections of Alzheimer's Disease Patients" by Gieffers
et al. Since we were the first to report on this issue (
Balin,
et al., 1998), I wanted to address how the present report differs quite
substantially from that which we reported.
My comments are outlined below:
The recent report by Gieffers,
et al. does not address our previous research findings using any of
the same protocols with which we made our findings. This point is acknowledged
by Gieffers et al. in their discussion of results. In fact, another report
that recently appeared in the literature (Nochlin,
et al., 1999) used similar protocols as did Gieffers et al and also
failed to detect C. pneumoniae in the Alzheimer's brain. In neither report
did the investigators try to replicate our findings in the way that was
clearly outlined in our 20-page report.
Given that such a different technical approach was taken, we can only
offer explanations as to the difficulties one encounters in using these
different approaches. In fact, we attempted to maximize our own technical
capabilities to address a very difficult question in identifying and localizing
an intracellular organism in the Alzheimer brain. For this reason and because
of the body of literature from hundreds of different Chlamydia studies that
demonstrate a tremendous variability in diagnosing, identifying, and localizing
Chlamydia in clinical specimens, we went to the extreme in our investigations
by using a large variety of techniques on different clinical specimens.
These techniques included: (A) both PCR and RT-PCR for genetic analysis
of C. pneumoniae from FROZEN tissues , not formalin-fixed paraffin embedded
tissues; (B) immunohistochemistry on formalin-fixed paraffin-embedded tissues
of 7-10 micron thickness, not on sections as thin as 4 microns as outlined
in the present report; also antigen retrieval and antibody dilutions from
our report were more varied and we used some different antibodies than used
in the present report (we have found that different aliquots of the anti-Chlamydial
antibodies have to be carefully titrated to maximize the immunoreaction);
(C) electron and immuno-electron microscopy to identify the organisms in
cells and tissues; (D) culturing from AD human brain homogenates the Chlamydia
organisms into human monocytes. In the latest reports, electron microscopy,
immuno-electron microscopy, and culturing from brain tissues were not performed.
We are confident that organisms containing genetic sequences and expressing
C. pneumoniae-related antigens were present in the samples examined by our
group. Our findings hold such great implications as to how inflammation
in the AD brain may be triggered by infection with C. pneumoniae that we
must demand that studies to replicate and/or validate our first report should
be performed with the rigor and comparable techniques that will provide
data that can truly be compared and analyzed. We cannot stop short of intensifying
these studies and we must work to develop standard techniques that can be
applied by laboratories analyzing clinical specimens. To fall short of this,
as we believe that the present publications do, only further demonstrates
the technical difficulties found in the Chlamydiology arena!
The fundamental question is what role C. pneumoniae plays in the pathogenesis
of Alzheimer's disease. Others should be compelled to address this question
as well. We are sure that the original observations of C. pneumoniae in atherosclerosis
were received skeptically, but this organism is now being fully investigated
for its role in that disease. We hope that similar investigations will be
forthcoming in AD as well because, as most objective scientists would agree,
we are far from understanding the basis for neurodegenerative processes
as they may relate to infectious agents such as Chlamydia, Borrelia, Herpes
simplex virus type I, and HIV.
The following additional paragraphs further discuss specifically why
we believe that the techniques used in the two reports which could not detect
C. pneumoniae in AD brains were not optimal for this determination. The tissues
used in the present article for PCR analysis were formalin-fixed and paraffin-embedded
tissue samples. In our PCR analysis, we used frozen tissue samples to minimize
the difficulties encountered in PCR from fixed and embedded tissues. In
particular, during extraction and re-hydration of paraffin-embedded fixed
tissue, fragmentation of the bacterial DNA may prevent amplification of
target sequences of > 400 bases so that nested PCR would not detect the
organism and the sensitivity of the technique does not apply due to the
starting material. In the present study, the authors state that "Successful
DNA extraction was ensured, since the PCR protocol used was previously evaluated
for vascular tissue and proven to be substantially more sensitive than cell
culture." This assumption is not necessarily valid when applied to
extraction of formalin-fixed paraffin-embedded brain tissues. The reasons
would include: that the organism presence and total bacterial burden in
brain tissues may vary quite extensively from those present in systemic
vascular tissues (eg, tissue specific tropism by a strain variant), that
the organism in brain tissues may have undergone forms of degradation compromising
the status of its DNA, and the copy number of the genes of interest may
be very low. In fact, in an article in J Clin Micro (36:1512-1517,
1998), in which Marchetti et al. evaluated PCR in the detection of another
organism from formalin-fixed, paraffin-embedded tissues, the authors came
to the conclusion that the efficacy of PCR strictly depends on several amplification
parameters which include DNA concentration, target DNA size, and the repetitiveness
of the amplified sequence. Furthermore, in the present study, extraction
of DNA with phenol-chloroform was described as using standard protocols.
I have two problems with this description. First, there is no reference
given here as to which standard protocol was used (there are numerous and
varied protocols that can be used); second, one cannot assume that the extraction
protocol was sufficient to extract the target DNA from C. pneumoniae organisms,
even though it apparently was sufficient for extracting eukaryotic DNA.
In the present study, the investigators used PCR primer sets that have
been shown to amplify C. pneumoniae genes in other types of clinical samples
(i.e., coronary artery tissues). However, these primer sets were not comparable
to what we used in our original study, and they may or may not work on C. pneumoniae in brain samples. The difficulties that many laboratories have
had in the PCR analysis of Chlamydia in numerous clinical samples, anywhere
from cardiovascular tissues, arthritis tissues, lung tissues, etc., indicates
that no standard PCR methodology has been successfully developed to detect
genetic material correlating to infection with Chlamydia pneumoniae. Given
the recognition of these difficulties by the investigators in the discussion
section of their present report, I find it curious as to why analysis of
Alzheimer samples was not performed in a manner that would at least approach
the protocol that we used in our original studies.
With regard to the immunocytochemistry experiments on which the present
study reports, we can make the following observations:
A. Our report used 7-10 micron thick formalin-fixed paraffin-embedded
sections for immunocytochemistry experiments, whereas the present report
used 4 micron thick sections. In our experience with C. pneumoniae infection
in the AD brain, we believe that it is best to use 7-10 micron thick sections
from paraffin-embedded tissues to obtain sections that include C. pneumoniae
inclusions within glial cells. Given the typical deparaffinization protocols
and rehydration, we believe that very thin paraffin sections are more friable
and cells that contain C. pneumoniae inclusions are even more suspect for
maintenance of integrity. Our tissue culture data supports this latter contention,
in that the cytoplasm and cytoplasmic vacuoles of C. pneumoniae-infected
cells is easily extruded following cytospinning at very low speeds (500-800
rpm).
B. Our report used antigen retrieval prior to primary antibody incubations;
this report does not mention antigen retrieval methods.
C. The antibodies used in our report included one (RR-402, Washington
Research Foundation) of which was used in the present report. We used dilutions
of 1:50 - 1:250 for this antibody and obtained our best results with the
dilutions at 1:50. Intriguingly, we also used the DAKO distributed antibody
from the same source (Wash. Res. Found.), but needed to use this lot at
1:5 dilution. We believe that different lots of the RR-402 antibody have
different antibody concentrations, and presently, must titrate for each
lot. This discrepancy may account for both positive and negative results
depending on the lot used. Curiously, in a different report (in Neurology
53:1888, 1999) by individuals from the University of Washington (location
of Washington Research Foundation), Department of Pathology and Pathobiology,
this antibody was not even used in their study that also failed to detect
C. pneumoniae in Alzheimer's brain tissues. We find their failure to use, basically,
the antibody to be remarkable given the association of this group with the
Washington Research Foundation. More intriguingly, this latter group of
investigators also used methods of immunocytochemistry and PCR that were
shown to be successful for detection of C. pneumoniae in atheromatous arterial
tissues (Kuo
et al., 1993), similar to that used by Gieffers et al. in the present
J Clin Micro report. I am addressing this issue here because in both reports
by Gieffers et al. and Nochlin et al. (the Neurology report cited above),
neither used methodology clearly outlined in our manuscript in Med. Micro.
Immuno. 187:23-42, 1998.
It is our belief that our thorough report using PCR, RT-PCR, immunocytochemistry,
electron microscopy, immunoelectron microscopy, and culture analysis, along
with all proper controls, should at least be mirrored in a comparable study
to obtain results that could be realistically compared for their techniques
and any discrepancies that may or may not be found. Unfortunately, the two
studies that have been performed have assumed that techniques found successful
for other tissue samples could be applied to brain samples that were formalin-fixed
and paraffin-embedded. Therefore, in our estimation, these studies are comparing
apples to oranges, and in essence they are just reaffirming the technical
difficulties and absence of standardization of techniques that are used
throughout the field of Chlamydiology and for application to clinical samples.
View all comments by Brian Balin
Home
